[XeTeX] Redefining \em and \emph

[Will Robertson]

On 2010-09-25 09:33:20 +0930, Gareth Hughes 
<garzohugo at gmail.com> said:

> Will Robertson wrote:
>> It doesn't seem to. In that case I'd write something like
>> 
>> \let\oldemph\emph
>> \renewcommand\emph{%
>> \if at RTL \expandafter\aemph \else \expandafter\oldelse \fi
>> }
> 
> I still can't figure out how to do this for \eminnershape as defined in
> fixltx2e. Any suggestions how to do that?

Do you mean so that you can write

    \aemph{ ... \aemph{ ...} ... }

and have the inner \aemph not put the overline on top? In this case, 
you can write (and this might be a reasonable addition to polyglossia)

\documentclass{article}
\makeatletter
\def\aemph#1{%
  \begingroup
    \let\aemph\inneraemph
    $\overline{\hbox{#1}}$%
  \endgroup
}
\def\inneraemph#1{%
  \egroup% close \hbox
  \egroup% close \overline
  $% close math
  {\eminnershape #1}% inner emphasis!
  $% "reopen" math
  \expandafter\overline\bgroup % "reopen" \overline
  \hbox\bgroup % "reopen" \hbox
}
\makeatother
\begin{document}
hello there \aemph{foo \aemph{oof rab} bar} baz black
\end{document}

Modulo checks for bidi and using \hboxR instead of \hbox and a 
"correct" definition for \eminnershape in this context, and so on. But 
this has nothing to do with the idea of \eminnershape: if you want to 
tie the two together, fontspec actually overrides fixltx2e's 
emph/eminnershape code:

\ExplSynaxOn
\DeclareRobustCommand \em {
  \@nomath\em
  \tl_if_eq:xxTF \f at shape \itdefault \eminnershape
  {
    \tl_if_eq:xxTF \f at shape \sldefault \eminnershape \emshape
  }
}
\DeclareTextFontCommand{\emph}{\em}
\cs_set_eq:NN \emshape \itshape
\cs_set_eq:NN \eminnershape \upshape
\ExplSyntaxOff

Not sure if that helps directly...

Will